| Message ID | 20240829185830.76913-1-vinicius.peixoto@canonical.com |
|---|---|
| Headers | show |
| Series | Fix race of mana_hwc_post_rx_wqe and new hwc response | expand |
On Thu, Aug 29, 2024 at 03:58:27PM -0300, Vinicius Peixoto wrote: > BugLink: https://bugs.launchpad.net/bugs/2078001 > > [Impact] > > Microsoft has requested us to backport the following bugfix, currently in the netdev tree: > > 8af174ea863c net: mana: Fix race of mana_hwc_post_rx_wqe and new hwc response > https://git.kernel.org/pub/scm/linux/kernel/git/netdev/net.git/commit/?id=8af174ea863c72f25ce31cee3baad8a301c0cf0f > > [Fix] > > Clean cherry-pick from the netdev tree to the {oracular,noble,jammy}:linux-azure trees. > > [Test plan] > > Build and boot tested. > > [Where problems could occur] > > This seems to have a low risk of regression, since it's just making sure to finish posting an RX workqueue event before waiting for completion in order to avoid a race condition with the hardware. However, any regressions here would likely impact users as MANA-related errors/timeouts. > > [Other info] > > SF #00393766 > > Haiyang Zhang (1): > net: mana: Fix race of mana_hwc_post_rx_wqe and new hwc response > > .../net/ethernet/microsoft/mana/hw_channel.c | 62 ++++++++++--------- > 1 file changed, 34 insertions(+), 28 deletions(-) Acked-by: Aaron Jauregui <aaron.jauregui@canonical.com>
Acked-by: Thibault Ferrante <thibault.ferrante@canonical.com> On 29-08-2024 20:58, Vinicius Peixoto wrote: > BugLink: https://bugs.launchpad.net/bugs/2078001 > > [Impact] > > Microsoft has requested us to backport the following bugfix, currently in the netdev tree: > > 8af174ea863c net: mana: Fix race of mana_hwc_post_rx_wqe and new hwc response > https://git.kernel.org/pub/scm/linux/kernel/git/netdev/net.git/commit/?id=8af174ea863c72f25ce31cee3baad8a301c0cf0f > > [Fix] > > Clean cherry-pick from the netdev tree to the {oracular,noble,jammy}:linux-azure trees. > > [Test plan] > > Build and boot tested. > > [Where problems could occur] > > This seems to have a low risk of regression, since it's just making sure to finish posting an RX workqueue event before waiting for completion in order to avoid a race condition with the hardware. However, any regressions here would likely impact users as MANA-related errors/timeouts. > > [Other info] > > SF #00393766 > > Haiyang Zhang (1): > net: mana: Fix race of mana_hwc_post_rx_wqe and new hwc response > > .../net/ethernet/microsoft/mana/hw_channel.c | 62 ++++++++++--------- > 1 file changed, 34 insertions(+), 28 deletions(-) >
On 8/29/24 1:58 PM, Vinicius Peixoto wrote: > BugLink: https://bugs.launchpad.net/bugs/2078001 > > [Impact] > > Microsoft has requested us to backport the following bugfix, currently in the netdev tree: > > 8af174ea863c net: mana: Fix race of mana_hwc_post_rx_wqe and new hwc response > https://git.kernel.org/pub/scm/linux/kernel/git/netdev/net.git/commit/?id=8af174ea863c72f25ce31cee3baad8a301c0cf0f > > [Fix] > > Clean cherry-pick from the netdev tree to the {oracular,noble,jammy}:linux-azure trees. > > [Test plan] > > Build and boot tested. > > [Where problems could occur] > > This seems to have a low risk of regression, since it's just making sure to finish posting an RX workqueue event before waiting for completion in order to avoid a race condition with the hardware. However, any regressions here would likely impact users as MANA-related errors/timeouts. > > [Other info] > > SF #00393766 > > Haiyang Zhang (1): > net: mana: Fix race of mana_hwc_post_rx_wqe and new hwc response > > .../net/ethernet/microsoft/mana/hw_channel.c | 62 ++++++++++--------- > 1 file changed, 34 insertions(+), 28 deletions(-) > Patches applied to oracular/noble/jammy:linux-azure master-next branches. Thanks, John
BugLink: https://bugs.launchpad.net/bugs/2078001 [Impact] Microsoft has requested us to backport the following bugfix, currently in the netdev tree: 8af174ea863c net: mana: Fix race of mana_hwc_post_rx_wqe and new hwc response https://git.kernel.org/pub/scm/linux/kernel/git/netdev/net.git/commit/?id=8af174ea863c72f25ce31cee3baad8a301c0cf0f [Fix] Clean cherry-pick from the netdev tree to the {oracular,noble,jammy}:linux-azure trees. [Test plan] Build and boot tested. [Where problems could occur] This seems to have a low risk of regression, since it's just making sure to finish posting an RX workqueue event before waiting for completion in order to avoid a race condition with the hardware. However, any regressions here would likely impact users as MANA-related errors/timeouts. [Other info] SF #00393766 Haiyang Zhang (1): net: mana: Fix race of mana_hwc_post_rx_wqe and new hwc response .../net/ethernet/microsoft/mana/hw_channel.c | 62 ++++++++++--------- 1 file changed, 34 insertions(+), 28 deletions(-)